DOI: 10.1126/science.1213397, 570 (2012);335 Science

, et al.Lisa C. KannerDuring the Last GlacialHigh-Latitude Forcing of the South American Summer Monsoon

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

): February 6, 2012 www.sciencemag.org (this infomation is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/335/6068/570.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2012/01/12/science.1213397.DC1.html can be found at: Supporting Online Material

http://www.sciencemag.org/content/335/6068/570.full.html#relatedfound at:

can berelated to this article A list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/335/6068/570.full.html#ref-list-1, 9 of which can be accessed free:cites 47 articlesThis article

http://www.sciencemag.org/content/335/6068/570.full.html#related-urls1 articles hosted by HighWire Press; see:cited by This article has been

http://www.sciencemag.org/cgi/collection/atmosAtmospheric Science

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2012 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Febr

uary

6, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

makes catalyst recovery challenging. We focusedour experiments on the cross-linking of two com-mercial silicone fluids: SilForce SL6100 andSilForce SL6020 (Momentive Performance Mate-rials,Waterford, NY) (Fig. 3). Addition of a toluenesolution containing 1000 ppm of (iPrPDI)Fe(N2)2to a neat mixture of the two fluids at 23°C resultedin immediate cross-linking.Dissolving the iron com-pound in a small amount of toluene is necessarybecause of the low solubility of (iPrPDI)Fe(N2)2in the silicone fluids. We also observed immedi-ate cross-linking upon addition of 500 ppm of[(MePDI)Fe(N2)]2(m2-N2) in toluene solution.Withthe more active [(MePDI)Fe(N2)]2(m2-N2) com-pound, cross-linking could be achieved over thecourse of 2 hours in neat silicone fluid withoutthe addition of toluene. In each case thatwe studied,the resulting silicone polymer was identical to ma-terial prepared commercially using platinum cat-alysts. Photographs are presented in Fig. 3 forvisual comparison.

Collectively, these results demonstrate thepotential of Earth-abundant metal complexesto compete with more rare and more expensiveprecious-metal counterparts. Beyond their highactivities, the iron compounds offer functional-group tolerance and, perhaps most importantly,exclusive regioselectivity that obviates the needfor separation of unwanted by-products obtainedin industrial processes.

References and Notes1. B. Marciniec, J. Gulinski, W. Urbaniak, Z. W. Kornetka,

in Comprehensive Handbook on Hydrosilylation,B. Marciniec, Ed. (Pergamon, Oxford, 2002), pp. 3–7.

2. B. Marciniec, Coord. Chem. Rev. 249, 2374 (2005).3. V. B. Pukhnarevitch, E. Lukevics, L. I. Kopylova,

M. Voronkov, Perspectives of Hydrosilylation (Institute forOrganic Synthesis, Riga, Latvia, 1992).

4. I. Ojima, in The Chemistry of Organic Silicon Compounds,S. Patai, Z. Rappoport, Eds. (Wiley Interscience, NewYork, 1989), pp. 1479–1526.

5. A. K. Roy, Adv. Organomet. Chem. 55, 1 (2007).6. A. F. Noels, A. J. Hubert, in Industrial Applications of

Homogeneous Catalysis, A. Mortreux, Ed. (Kluwer,Amsterdam, 1985), pp. 80–91.

7. E. Plueddemann, Silane Coupling Agents (Plenum Press,New York, ed. 2, 1991).

8. R. M. Hill, in Silicone Surfactants, Surfactants ScienceSeries, vol. 86, R. M. Hill, Ed. (Marcel Dekker, New York,1999), pp. 1–48.

9. M. Wong, E. Memisha, U.S. Patent 6,805,856 (2004).10. I. E. Markó et al., Science 298, 204 (2002).11. O. Buisine et al., Chem. Commun. 2005, 3856 (2005).12. P. B. Hitchcock, M. F. Lappert, N. J. W. Warhurst,

Angew. Chem. Int. Ed. Engl. 30, 438 (1991).13. A. J. Holwell, Platin. Met. Rev. 52, 243 (2008).14. C.-J. Yang, Energy Policy 37, 1805 (2009).15. P. J. Chirik, K. Wieghardt, Science 327, 794 (2010).16. S. Enthaler, K. Junge, M. Beller, in Iron Catalysis in

Organic Chemistry, B. Plietker, Ed. (Wiley-VCH,Weinheim, Germany, 2008), pp. 125–142.

17. W. M. Czaplik, M. Mayer, J. Cvengros, A. J. von Wangelin,ChemSusChem 2, 396 (2009).

18. J. Y. Corey, J. Braddock-Wilking, Chem. Rev. 99, 175(1999).

19. M. Zhang, A. Zhang, Appl. Organomet. Chem. 24, 751(2010).

20. M. A. Schroeder, M. S. Wrighton, J. Organomet. Chem.128, 345 (1977).

21. R. D. Sanner, R. G. Austin, M. S. Wrighton, W. D. Honnick,C. U. Pittman, Inorg. Chem. 18, 928 (1979).

22. R. H. Morris, Chem. Soc. Rev. 38, 2282 (2009).23. P. J. Chirik, in Catalysis Without Precious Metals

(Wiley-VCH, Weinheim, Germany, 2010), pp. 83–106.24. B. Marciniec, Silicon Chem. 1, 155 (2002).25. J. C. Mitchener, M. S. Wrighton, J. Am. Chem. Soc. 103,

975 (1981).26. F. Kakiuchi, Y. Tanaka, N. Chatani, S. Murai, J. Organomet.

Chem. 456, 45 (1993).27. S. C. Bart, E. Lobkovsky, P. J. Chirik, J. Am. Chem. Soc.

126, 13794 (2004).28. J. Y. Wu, B. N. Stanzl, T. Ritter, J. Am. Chem. Soc. 132,

13214 (2010).29. S. K. Russell, J. M. Darmon, E. Lobkovsky, P. J. Chirik,

Inorg. Chem. 49, 2782 (2010).30. J. W. Sprengers, M. de Greef, M. A. Duin, C. J. Elsevier,

Eur. J. Inorg. Chem. 2003, 3811 (2003).31. R. J. Trovitch, E. Lobkovsky, P. J. Chirik, J. Am. Chem. Soc.

130, 11631 (2008).32. P. J. G. Stevens, Pestic. Sci. 38, 103 (1993).

Acknowledgments: We thank Momentive PerformanceMaterials for financial support. Portions of this work have beenpublished in the following patent applications: US2011/0009573A1 and U.S. serial no. 13/325,250.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/335/6068/567/DC1SOM TextFigs. S1 to S14Tables S1 to S3References (33, 34)

26 September 2011; accepted 29 November 201110.1126/science.1214451

High-Latitude Forcing of theSouth American Summer MonsoonDuring the Last GlacialLisa C. Kanner,1* Stephen J. Burns,1 Hai Cheng,2,3 R. Lawrence Edwards3

The climate of the Last Glacial period (10,000 to 110,000 years ago) was characterized by rapidmillennial-scale climate fluctuations termed Dansgaard/Oeschger (D/O) and Heinrich events.We present results from a speleothem-derived proxy of the South American summer monsoon(SASM) from 16,000 to 50,000 years ago that demonstrate the occurrence of D/O cycles andHeinrich events. This tropical Southern Hemisphere monsoon reconstruction illustrates anantiphase relationship to Northern Hemisphere monsoon intensity at the millennial scale. Ourresults also show an influence of Antarctic millennial-scale climate fluctuations on the SASM.This high-resolution, precisely dated, tropical precipitation record can be used to establish thetiming of climate events in the high latitudes of the Northern and Southern Hemispheres.

Dansgaard/Oeschger (D/O) and Heinrichevents have been recognized in manyterrestrial and marine records, particular-

ly from the Northern Hemisphere. First observedin oxygen isotope ratios of ice cores from Green-

land (1), the D/O cycles are interpreted as largeabrupt warmings (interstadials) followed bymoregradual cooling to stadial conditions. The D/Ocycles are particularly well expressed in tropicalNorthern Hemisphere speleothem records, wherecold and warm periods in Greenland are coin-cident with dry and wet periods, respectively,in the East Asian and Indian summer monsoons(EASM and ISM) (2–4). The presence of D/Ocycles in the Southern Hemisphere, or their rela-tion to Southern Hemisphere monsoons, has notbeen established, however. Heinrich events arepronounced cold intervals (~500 T 250 years in

duration) marked by horizons of ice-rafted debrisin ocean sediment cores from mid-latitudes inthe Atlantic Ocean (5–7). Although an interhemi-spheric antiphased relation has been observed forthe Heinrich (H1 to H6) events, the specific tim-ing and duration are less well constrained be-cause of uncertainties in ocean reservoir ages (8)and changes in sediment accumulation rates. Herewe present a high-resolution, precisely dated, speleo-them reconstruction of South American summermonsoon (SASM) intensity for the Last Glacialperiod. This record demonstrates the presenceof D/O cycles in the SASM, where an intensifiedSASM is associated with stadial events in Green-land. In addition, we identify millennial-scale cli-mate events that resulted from the interplay betweenrapid Antarctic warming and cool North AtlanticHeinrich events. Finally, we use the signature ofthe major D/O cycles to propose possible con-straints on the timing and duration of Heinrichevents H4 and H5.

We collected stalagmite P09-PH2 fromPacupahuain Cave (11.24°S, 75.82°W; 3800 mabove sea level) in the central Peruvian Andes(fig. S1). Pacupahuain Cave was formed in aTriassic dolomitic limestone massif (9) and in anarea where glacial landforms are not present.P09-PH2 is a calcite stalagmite 16 centimeterstall and was collected in the cave’s main galleryabout 500 m from the entrance. The speleothemwas halved along the growth axis and subsam-pled along growth layers for radiometric dating,using uranium-thorium (U-Th) techniques by

1Department of Geosciences, University of Massachusetts,Amherst, MA 01002, USA. 2Institute of Global EnvironmentalChange, Xi'an Jiaotong University, Xi'an 710049, China. 3De-partment of Geology and Geophysics, University of Minnesota,Minneapolis, MN 55455, USA.

*To whom correspondence should be addressed. E-mail:lkanner@geo.umass.edu

3 FEBRUARY 2012 VOL 335 SCIENCE www.sciencemag.org570

REPORTS

on

Febr

uary

6, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

multicollector inductively coupled plasma massspectroscopy (MC-ICP-MS) (10, 11). Seventeen230Th dates, all in stratigraphic order, have ana-lytical errors <0.4% (fig. S2 and table S1). An agemodel was developed based on linear inter-polation between adjacent age measurements,yielding an age range for the sample of 16,020to 49,470 years before the present (yr B.P., withthe present = 1950).

Stable oxygen and carbon isotope ratios(d18O and d13C) were measured on 1076 micro-milled samples taken along the growth axis. Theaverage sampling resolution is approximately

30 years. The very low correlation (coefficientof determination r2 = 0.1) between d13C and d18Ovalues along the growth axis (fig. S3) indicatesthat kinetic fractionation did not significantly in-fluence speleothem isotopic variability (11, 12).P09-PH2 d18O values range from –14 per mil(‰) to –17.5‰ between 16 and 50 thousand yrB.P. (ky B.P.) (Fig. 1). The d18O record is charac-terized by high-amplitude millennial-scale events,with rapid increases of up to 2.5‰ occurringover a century or less. The decline to local d18Ominima occurred more gradually, over periodsranging from a few centuries to a few millennia.

The lowest values in the record occur over anextended period of time between 28.8 to 30.3 kyB.P. The highest values are attained over verybrief intervals at 16, 37, 38, and 47 ky B.P.

We interpret the changes in speleothem cal-cite to be primarily driven by changes in the iso-topic composition of precipitation at the site. Twoadditional factors, changes in cave temperatureand changes in the isotopic composition of sea-water, also contribute to the observed variability,but serve primarily to reduce the amplitude ofisotopic change by a maximum of about 1.5‰(11). Thus, P09-PH2 records the minimum am-plitude of changes in d18O of precipitation, andthe actual changes were probably larger. What,then, drives changes in d18O of precipitation atthe study site?

The primary climatic control on rainfall d18Ofor tropical South America is the amount of pre-cipitation (13). Consistent with the “amount ef-fect” (14) and Rayleigh-type fractionation duringrainout, a stronger SASM leads to lower d18Ovalues in tropical South American rainfall (15).Similar correlations have been made for otherparts of the tropical Andes, where local precip-itation d18O is strongly anticorrelated to rainfallamount upstream in the Amazon Basin rather thanat the study site (16, 17). Rayleigh distillation andmoisture recycling in the Amazon Basin (18) pre-serve the isotopic signal of the monsoon alongthe moisture path from eastern Amazonia up tothe central Peruvian Andes (15). On interannualtime scales, the SASM is influenced by sea sur-face temperatures (SSTs) in the Atlantic and Pa-cific Oceans (19) and on orbital time scales bychanges in insolation (20, 21). These studies dem-onstrate that rainfall d18O in the central Andesrecords changes in the intensity of large-scale con-tinental atmospheric convection, with more neg-ative rainfall d18O values being indicative ofenhanced SASM activity and increased rainout.

Comparison of themillennial-scale variabilityin the Pacupahuain Cave record to the isotopicfluctuations from Greenland (Fig. 1) shows thatall D/O cycles are also found in our record ofSASM intensity, with the sole exception of D/Oevent 3, which for unknown reasons is not present.The identification of D/O events is based on com-parison to the North Greenland Ice Core Project(NGRIP) oxygen isotope record, using the ice coreage model of (22) (Fig. 1) and the MSD speleo-them record fromHuluCave, China, using its ownprecise chronology (fig. S4) (4, 11). Abrupt in-creases in NGRIP d18Oice, which characterize theonset of the D/O events, are clearly associated withrapid increases in d18Ocalcite for P09-PH2. The re-lationship betweenD/O events in theNGRIP recordand our results demonstrates that warm events inGreenland correspond to intervals of decreasedSASM intensity. The timing of millennial-scalevariability is synchronous within the independentage model errors.

Figure 1 also includes the record of atmospher-ic methane (CH4) concentrations from Green-land and Antarctic ice cores (23–25). The origin

Fig. 1. The d18O reconstruction of stalagmite P09-PH2 compared to high-latitude ice core d18O and CH4concentrations. SMOW, standard mean ocean water standard; ppbv, parts per billion by volume; VPDB,Vienna Pee Dee belemnite standard. (A) GRIP and NGRIP CH4 (black) (24, 25) and EDML CH4 (purple) (23)concentrations. (B) NGRIP ice core d18O (dark green, 7-point runningmean) (22). (C) Speleothem P09-PH2d18O time series with U/Th dates with 2s uncertainty bars (dark blue, 3-point running mean). (D) EDML icecore d18O (red, 7-point running mean) (23). The age models for Greenland and Antarctica d18O and NGRIPand EDML CH4 concentrations from (22) are shown. GRIP CH4 concentrations are shown on the GreenlandIce Sheet Project 2 time scale from (25). The time scale of P09-PH2 is given on its own age model (fig. S2).The Heinrich events are depicted by the vertical arrows, and age estimations are from (4, 7). The numbersindicate Greenland interstadials and Antarctic AIM events.

www.sciencemag.org SCIENCE VOL 335 3 FEBRUARY 2012 571

REPORTS

on

Febr

uary

6, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

of millennial-scale variations in atmosphericCH4 is not certain, although they are thought toarise from changes in the flux from global wet-lands (26, 27) rather than from the release of CH4

from marine clathrates (28) or changes in CH4

removal via oxidation in the troposphere (26).Whether source flux changes derive from tropicalwetlands or boreal forests has not been established.We observe that increases in global atmosphericCH4 occur contemporaneously with increasesin P09-PH2 d18O and reduced SASM intensity(Fig. 1). Thus, the Southern Hemisphere SouthAmerican tropics are unlikely to be a source ofincreased atmospheric CH4, via wetland expan-sion, during Greenland interstadials. The inter-hemispheric antiphasing of the tropical monsoonsystems during D/O events, however, indicatesthat rapid northward shifts of the Intertropical Con-vergence Zone (ITCZ) and intensification of theEASM and ISM might lead to wetland expan-sion and increased CH4 production in the North-ern Hemisphere tropics during warm Greenlandintervals.

In addition to themillennial-scale D/O events,each Heinrich event (H1 through H5) can also beidentified in the P09-PH2 reconstruction by peri-ods of enhanced SASM activity (Fig. 1) and bycorrelation to decreased Asian monsoon intensitybased on the Hulu Cave record (fig. S4) (4, 11).Previous paleoclimate studies of the SASM,which reconstructed precipitation intensity dur-ing the Last Glacial period but at lower resolutionthan presented here, also identified Heinrich eventsas particularly wet intervals (29, 30). IncreasedHeinrich event precipitation over tropical SouthAmerica is supported by evidence of substantiallateral advancement of central Andean glaciersduring H1 and H2 (31) and increases in pre-cipitation over northeast Brazil for H1 throughH5 (30, 32, 33).

The expression of D/O cycles and Heinrichevents in SASM variability is antiphased withNorthern Hemisphere monsoon precipitationreconstructions over the Last Glacial period. Ourrecord indicates a weak SASM during D/O inter-stadials, but reconstructions of Indian Ocean cli-mate (2) and EASM intensity (4) demonstratevery wet intervals for interstadial phases. Althoughprevious work has demonstrated such antiphasebehavior for orbital-scale climate variability in thetropics (21) and for Heinrich events (30), it hadnot been unequivocally determined for D/O scalevariability. Such antiphase behavior can be ex-plained by latitudinal shifts of the ITCZ drivenby rapid Northern Hemisphere temperature fluc-tuations.Modeling data demonstrate ITCZmigra-tion away from the hemisphere with high-latitudecooling and added ice cover (34). During pro-nounced North Atlantic cold events, southwardmigration of the ITCZ can alter Hadley cell cir-culation, increase subsidence (dry conditions)for the northern branch, and intensify uplift (wetconditions) of the southern branch (35, 36). Asa result, the ISM and EASM weaken and theSASM intensifies. Although our record is short

relative to precessional-scale insolation changes,insolation appears to be only weakly expressed inour record (fig. S5).

In addition to demonstrating the presence ofD/O and Heinrich variability in the SASM, ourrecord reveals additional rapid, millennial-scaleclimate reversals in the tropical climate record.These events appear to arise from the interplaybetweenNorthern and SouthernHemisphere high-latitude climate change. The reversals are mostclearly expressed at 38.75 and 48 ky B.P. duringextended Greenland stadials (Fig. 2). We proposethat the onset of these events coincides with theonset of Antarctic Isotope Maxima (AIM) events8 and 12 in the EPICA (European Project forIce Coring in Antarctica) DronningMaud Land(EDML) ice core record (22). Prolonged Green-land cold phases have been linked with inten-sification and southward displacement of theSouthern Hemisphere westerlies, enhanced up-welling in the SouthernOcean, rising atmosphericCO2, and Antarctic warming (37, 38). Specifi-cally, the opal flux reconstruction presented in(38) demonstrates that ocean upwelling and CO2

rises were largest during H4 and H5 as comparedtomore recent Heinrich events. Thus, we propose

that enhanced changes in Southern Ocean cli-mate specific to H4 and H5 can explain why ourrecord shows reversals in SASM vigor at H4 andH5 and not for the later Heinrich events.

The precise relation between these transientmillennial-scale events and the ice-rafted com-ponents of Heinrich events H4 and H5, whichoccurred at approximately the same time period,is less clear. It is well established that the endingsof H4 and H5 occur with Greenland warming atD/O events 8 and 12, respectively, but the be-ginnings of H4 and H5 are less well constrained.One interpretation based on our record is that theonset of Heinrich events H4 and H5 began withthe sharp increase in SASM intensity that followsthe decrease associated with Antarctic warming.An abrupt increase in SASM intensity associatedwith Heinrich events has been recognized inmarine (32, 33) and terrestrial (29, 30) records.This sequence of events would restrict the dura-tion of H4 and H5 events to <400 years, whichhas previously been suggested for H4 (39). Itwould also indicate that Antarctic warmingprecedes the start of Heinrich events 4 and 5.Antarctic warming has previously been proposedas a trigger for H1 (40). A second possibility,

Fig. 2. Detailed view of the d18O records from Pacupahuain Cave (blue) and Greenland (green) andAntarctic (red) ice cores during (A) Heinrich event 4 and D/O and AIM events 8 and (B) Heinrich event 5and D/O and AIM events 12. The solid black vertical line indicates the end of Heinrich events H4 and H5,and the dashed black vertical lines indicate two possible onsets for Heinrich events H4 and H5. Isotoperecords and corresponding age models are the same as in Fig. 1.

3 FEBRUARY 2012 VOL 335 SCIENCE www.sciencemag.org572

REPORTS

on

Febr

uary

6, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

however, is that H4 and H5 may have begun be-fore the weakening of the SASM and Antarcticwarming and thus lasted throughout the extendedGreenland cold phase. Although in this case wehave no explanation for why the SASM suddenlystrengthened.

One final aspect of the SASM reversalsbefore D/O events 8 and 12 is unusual. Highnorthern latitude warming is associated with anintensification of the Northern Hemisphere mon-soons. By analogy, high southern latitude warm-ing might be expected to result in intensificationof the SASM, but the opposite is observed. Thisseeming contradiction can be reconciled by com-parison of our record to a high-resolution SSTrecord based on alkenone paleothermometry fromthe Bermuda Rise (33.69°N, 57.56°W) (41). CoreMD95-2036 shows large, rapid, warm intervals(warmer by 3° to 5°C) immediately before D/Oevents 8 and 12, which were originally interpretedas instabilities during the rapid transitions of thelargest Greenland interstadials. Based on visualcomparison, we establish tie points between theBermuda Rise SST record and our d18O recordand provide a revised age model for these twooscillations in Bermuda Rise SSTs (11) (Fig. 3).The original marine core chronology was basedon maximizing the correlation to the Greenlandice core record and its chronology at the time ofpublication (11, 42). Our revised chronology for

the two intervals of MD95-2036 places the warmintervals 600 T 100 years earlier, relative to anupdated Greenland chronology, which is wellwithin the thousand-year uncertainty of the orig-inal marine core agemodel. Thus, Antarctic warm-ing at AIM events 8 and 12 may be associatedwith a weakened SASM as described above (Fig.2), as well as a warmer subtropical North AtlanticOcean (Fig. 3) (41). Similar temperature rever-sals within H4 and H5 have been identified in adeep water record from the Iberian Margin (43).It is not clear why Antarctic warming is associ-ated with warming in the subtropical North At-lantic, although one result of the latter could be anorthward shift of the ITCZ in the Atlantic Oceansector (44) due to a weakening of the subtropicalAtlantic gyre. The association betweenwarmSSTsin the Bermuda Rise region, a weakening of theSASM, and amore northerly position of the ITCZ,as we have described for the D/O events, couldalso be applied to the additional reversals iden-tified in Fig. 3.

References and Notes1. W. Dansgaard et al., Nature 364, 218 (1993).2. S. J. Burns, D. Fleitmann, A. Matter, J. Kramers,

A. A. Al-Subbary, Science 301, 1365 (2003).3. Y. Wang et al., Nature 451, 1090 (2008).4. Y. J. Wang et al., Science 294, 2345 (2001).5. H. Heinrich, Quat. Res. 29, 142 (1988).6. G. Bond et al., Nature 360, 245 (1992).

7. S. R. Hemming, Rev. Geophys. 42, RG1005 (2004).8. T. F. Stocker, Radiocarbon 40, 359 (1998).9. E. J. Cobbing et al., in The Geology of the Western

Cordillera of Northern Peru (Great Britian Institute ofGeological Sciences, Natural Environment ResearchCouncil, London, 1981), p. 143.

10. H. Cheng et al., Geology 37, 1007 (2009).11. Information on materials and methods is available as

supporting material on Science Online.12. C. H. Hendy, Geochim. Cosmochim. Acta 35, 801 (1971).13. M. Vuille, R. S. Bradley, M. Werner, R. Healy, F. Keimig,

J. Geophys. Res. 108, 4174 (2003).14. W. Dansgaard, Tellus 16, 436 (1964).15. M. Vuille, M. Werner, Clim. Dyn. 25, 401 (2005).16. F. Vimeux, R. Gallaire, S. Bony, G. Hoffmann, J. Chiang,

Earth Planet. Sci. Lett. 240, 205 (2005).17. G. Hoffmann et al., Geophys. Res. Lett. 30, 1179

(2003).18. R. L. Victoria, L. A. Martinelli, J. Mortatti, J. Richey, Ambio

20, 384 (1991).19. M. Vuille, R. S. Bradley, F. Keimig, J. Geophys. Res. 105,

12447 (2000).20. G. Seltzer, D. Rodbell, S. Burns, Geology 28, 35

(2000).21. F. W. Cruz Jr. et al., Nature 434, 63 (2005).22. B. Lemieux-Dudon et al., Quat. Sci. Rev. 29, 8 (2010).23. C. Barbante et al.; EPICA Community Members, Nature

444, 195 (2006).24. J. Flückiger, Global Biogeochem. Cycles 18, GB1020

(2004).25. T. Blunier, E. J. Brook, Science 291, 109 (2001).26. J. Chappellaz et al., Nature 366, 443 (1993).27. A. Dällenbach et al., Geophys. Res. Lett. 27, 1005

(2000).28. M. Bock et al., Science 328, 1686 (2010).29. X. Wang et al., Quat. Sci. Rev. 25, 3391 (2006).30. X. F. Wang et al., Nature 432, 740 (2004).31. J. A. Smith, D. T. Rodbell, J. Quat. Sci. 25, 243

(2010).32. A. Jaeschke, C. Ruhlemann, H. Arz, G. Heil, G. Lohmann,

Paleoceanography 22, PA4206 (2007).33. T. C. Jennerjahn et al., Science 306, 2236 (2004).34. J. C. H. Chiang, C. M. Bitz, Clim. Dyn. 25, 477 (2005).35. R. Zhang, T. L. Delworth, J. Clim. 18, 1853 (2005).36. A. C. Clement, L. C. Peterson, Rev. Geophys. 46, RG4002

(2008).37. G. H. Denton et al., Science 328, 1652 (2010).38. R. F. Anderson et al., Science 323, 1443 (2009).39. D. Roche, D. Paillard, E. Cortijo, Nature 432, 379

(2004).40. A. J. Weaver, O. A. Saenko, P. U. Clark, J. X. Mitrovica,

Science 299, 1709 (2003).41. J. P. Sachs, S. J. Lehman, Science 286, 756 (1999).42. E. A. Boyle, Proc. Natl. Acad. Sci. U.S.A. 94, 8300

(1997).43. L. C. Skinner, H. Elderfield, Paleoceanography 22,

PA1205 (2007).44. A. J. Broccoli, K. A. Dahl, R. J. Stouffer, Geophys. Res.

Lett. 33, L01702 (2006).

Acknowledgments: This work is supported by NSF grantATM-1003466 to S.J.B., NSF grants 0502535 and 1103403to R.L.E. and H.C., and Geological Society of America StudentAward 9308-10 to L.C.K. We thank C. Morales-Bermudez forhis invaluable assistance in the field. Comments from twoanonymous reviewers substantially improved the manuscript.The data in this paper are tabulated in the supporting onlinematerial and are archived online at ftp://ftp.ncdc.noaa.gov/pub/data/paleo.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/science.1213397/DC1Materials and MethodsFigs. S1 to S5Table S1References (45–51)

31 August 2011; accepted 4 January 2012Published online 12 January 2012;10.1126/science.1213397

Fig. 3. Bermuda Rise SSTs (orange) shown to depth in core MD95-2036 (41). D/O events wereidentified on the basis of visual correlation to the Greenland isotope temperature record as in (41). Anexpanded view of results from our P09-PH2 record (blue) and subtropical Atlantic SSTs (orange) usingour revised age model (11) is shown for the intervals of H4 (left) and H5 (right).

www.sciencemag.org SCIENCE VOL 335 3 FEBRUARY 2012 573

REPORTS

on

Febr

uary

6, 2

012

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m